Meaningful AI, explained with an umbrella

I keep coming across AI products that make me wonder whether anyone really needs them. The technology might be impressive, but the value isn’t always obvious.

I think design offers a useful way to understand how we can build more meaningful AI products. And we can explain it with something as ordinary as an umbrella.

I call it “The Umbrella Principle”.

About ten years ago, I studied industrial design at Design School Kolding. Recently, I revisited some of the articles we read back then. A central question in design is how we create value.

We can think of it as a simple equation:

WHAT we create + HOW it works = VALUE

An umbrella is a good example. The umbrella is what we create (WHAT). It shields us from the rain (HOW), so we can stay dry (VALUE).

To create that value, we need to work out the WHAT and the HOW. We can approach this in two ways.

Designing with a fixed HOW

Here, we have already decided HOW we will create value. The design task is to figure out WHAT to make.

If we have decided that we will keep people dry by shielding them from the rain, we have already narrowed the task. An umbrella or a raincoat becomes an obvious answer to WHAT we should create.

But other possibilities emerge when we approach the task differently.

Designing with an open HOW

The value we want to create stays the same, but we now have two unknowns. Both the WHAT and the HOW remain open.

This lets us ask different questions. Do we actually need to shield people from the rain? Could we help them avoid it by going outside at a different time?

Our new HOW could be to time a walk around the weather. Our WHAT could be a service that suggests when to leave to avoid the rain.

The value is still staying dry. We have simply found another way to get there.

What this means for AI

My impression is that many AI products start with the question, “How can we add AI?” That fixes the HOW before we have properly explored what we want to achieve. The design task then becomes finding a WHAT that justifies the choice.

If we start with the value we want to create and keep the HOW open, we give ourselves room to explore which approach makes sense. AI might turn out to be a brilliant part of the solution. But we need to do the work to find out.

Working with an open HOW is messier. You need to explore more possibilities, question your assumptions and be willing to change direction. Sometimes you have to admit that an idea has led you down a dead end.

A good designer knows how to make that mess productive. Small prototypes and tests help us gather feedback, understand what works and gradually find a way to create something useful.

AI is a fantastic tool. It can be an excellent answer to HOW when it makes sense for what we are trying to achieve. It can help us do things we couldn’t do before and create value we hadn’t imagined.

Those are the AI products I would like to see more of.